Perimeter computer networks, utilize various content filtering techniques to minimize unwanted electronic mail (“e-mail”) messages (i.e., “spam”) from being received by internal computer networks, such as an organization's mail system. Current content filtering techniques include filtering e-mail at the perimeter network by using algorithms to evaluate and assign “spam confidence levels” to incoming e-mail messages to identify spam. When a spam confidence level assigned to an incoming message reaches a configured threshold, the message is rejected. However, perimeter network content filtering algorithms which utilize spam confidence levels do not always accurately distinguish legitimate e-mail messages from spam thereby resulting in the occurrence of false positives (i.e., the identification of legitimate e-mail messages as spam). As a result, legitimate e-mail messages are prevented from being delivered to the internal network. Moreover, while more reliable content filtering techniques exist within internal networks, these internal network filtering techniques are not easily applied to the perimeter network. For example, safe recipients/blocked sender lists are typically only stored inside of internal networks and thus are unavailable for use by the perimeter network without engaging in the time consuming process of manually copying these lists from mailboxes associated with each user in the internal network. It is with respect to these considerations and others that the various embodiments of the present invention have been made.